Open calls for Short Term Scientific Missions

(Please click on the ‘green circle’ with ‘+’ sign for more info)
Contact person
Lab & Location
Available period
Institute & Department
Expertise
Andreas KasslerDISCO research group, Karlstad University, Sweden
jointly together with
VoerEir AB, Stockholm, Sweden
anytimeComputer Science Department
65188 Karlstad
Sweden
Performance Assessment of Cloud Infrastructures in relation to NFV Services

The most dramatic technology transformation in 130 years is starting in Telecom called Network Function Virtualization (NFV). This change is coinciding with Opensource becoming the new way of implementing cloud infrastructures with technologies such as Openstack, KVM, CEPH Opendaylight becoming the new standard. The telecom operators and suppliers are therefore facing a transforming Telecom Industry where new players and way of thinking is required. This change is not only happening in the Telecom industry but also in several adjacent verticals.

The proposed STSM will consist of two parts. The initial phase is focused on creating a theoretical model for potential “critical requirements” from a set of NFV applications on the underlying cloud/virtualization infrastructure as well as identifying open source tools (such as Kloudbuster) that can measure this requirement where possible. There should be two NFV services analysed that typically have complementary requirements in resource dimensions such as CPU or disk access. An example for a CPU intensive NFV service could be a virtualized router while an exampel for a disc intensive NFV service could be the VLR due to the frequent database access.

The second phase has the two objectives. The first objective is to understand how many of the potential “critical requirements” in the theoretical model from the initial phase can be measured and verified with the identified open source tools in practice. This is verified by conducting performance tests with the tools on different cloud infrastructures (or different configurations of a given Cloud Infrastructure like using different type of Kernel level offloading for packet processing). The second objective is to verify to what extent different levels/conditions in the specific “critical requirement” changes the performance of the application. This will be achieved by implementing in e.g. Kloudbuster mechanisms to put a given Virtual Machine that hosts the NFV service under CPU or disk stress and observe different Hypervisor metrics and correlate those metrics with performance indicators of the NFV service.

The outcome of this study will be of high interest for both Karlstad University and VoerEir AB. VoerEir AB is a newly started company placed in Stockholm(Sweden) with the mission to establish a unique market position and technology leadership in this transforming industry.
Dr. Reza MalekianSENTECH Chair in Broadband Wireless Multimedia Communication, University of Pretoria, Pretoria, Republic of South AfricaFeb 2015 to Dec 2015Department of Electrical, Electronic and Computer Engineering, University of Pretoria, Republic of South Africai) Internet of Things (Internet of Nano Things for healthcare, Internet of Things centric Cloud, Internet of Energy)
ii) Cloud computing and big data
iii) Mobile and wireless communications
iv) Advanced sensor networks

Tihana Galinac GrbacSoftware Engineering and Information Processing Laboratory
Rijeka, Croatia
Any timeSoftware Engineering and Information Processing Laboratory
Faculty of Engineering, University of Rijeka, Croatia
Prediction and performance, quality and reliability for composite services
Software engineering of large scale software systems and service chains
Anna BrunstromDistributed Systems and Communications Research Group (DISCO), Karlstad, SwedenAvailable immediatelyKarlstad University, Department of Mathematics and Computer ScienceThe proposed task for the STSM concerns development and evaluation of transport protocol mechanisms with a particular focus on cellular access. In particular we will investigate the potential of using metadata regarding e.g., content and congestion, when it comes to improving the end user experience in terms of less rebuffering, faster responses etc.

The candidate is expected to be familiar with network simulation, preferably ns3. He or she will implement an ns3 model of a transport protocol with metadata and evaluate different ways of using such data with respect to, e.g., flow prioritisation and congestion control in small but realistic scenarios. Own ideas in similar directions will be welcome. The candidate will be based at Karlstad University (KAU) and the work will be undertaken in cooperation with KAU and Ericsson. The work should result in an ns3 simulation framework and one or more high quality publication(s).
Philipp LeitnerZurich, SwitzerlandOn requestsoftware evolution & architecture lab, University of ZurichWe are interested in general topics of cloud- and service-based software engineering. An overview is available online on my web page: http://wp.ifi.uzh.ch/leitner/?page_id=98

Some current papers and preprints that give a good idea of what kind of research we like to work on:

Cloud benchmarking: http://arxiv.org/pdf/1411.2429v1.pdf
Empirical research / developer studies: http://arxiv.org/pdf/1409.6502v1.pdf
Middleware: http://arxiv.org/pdf/1411.2392v1.pdf
Software Evolution: https://mbrandt232.github.io/project/sqa-profiles/publication.pdf
Service Composition: http://www.infosys.tuwien.ac.at/staff/leitner/TUV-1841-2011-01.pdf
Steven LatreAntwerp, BelgiumAny timeUniversity of Antwerp - iMinds, BelgiumThe research focuses on cognitive network and service management in two main application domains:
1) Internet of Things centric cloud computing
2) Large-scale Wireless infrastructures: wireless sensor networks, WiFi hotspots, etc.
We mainly focus on the large-scale and chaotic environments and study aspects such as autonomic protocols, topology management, interference management, etc.
People who have expertise in AI techniques or similar (reinforcement learning, game theory, etc) are most welcome.

The topic of the STSM is something we would typically define together with the candidate. We do however have a concrete idea yet (but everything is up for discussion). We are doing a project with the largest dance festival in the world (Tomorrowland), where we are building a large-scale wireless sensor network. A STSM could aid in the validation of new solutions into a trial (an actual music festival).
Joel RodriguesCovilhã, PortugalAny time.Instituto de Telecomunicações, University of Beira Interior, PortugaleHealth, Mobile Health, and Ambient Assisted Living; Internet of Things; Vehicular Communications
Katarzyna WacCopenhagen, Denmark http://www.diku.dk/english/research/hcc/
OR Geneva, Switzerland http://www.qol.unige.ch/index.html
all except July-AugustUniversity of Copenhagen
Department of Computer Science
Denmark
Mobile communications and computing, wireless networking, network performance, user experience for mobile applications, user mobility models, context-awareness, Android OS programming expertise;

Access to mQoL Living Lab collecting mobile users data for 2+ years now from QoL lab (Uni Geneva, Switzerland).